tcg-ppc64: Use tcg_out_{ld,st,cmp} internally

Rather than using tcg_out32 and opcodes directly.  This allows us
to remove LD_ADDR and CMP_L macros.

Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ppc64: Relax register restrictions in tcg_out_mem_long

In order to be able to use tcg_out_ld/st sensibly with scratch
registers, assert only when we'd incorrectly clobber a scratch.

Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ppc64: Move functions around

Code movement only.  This will allow us to make use of the
other tcg_out_* functions in tidying their implementations.

Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ppc64: Avoid some hard-codings of TCG_TYPE_I64

Using more appropriate _PTR or _REG where possible.

Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ppc: Use uintptr_t in ppc_tb_set_jmp_target

Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>

Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging

Block pull request

# gpg: Signature made Mon 23 Jun 2014 09:53:49 BST using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"

* remotes/stefanha/tags/block-pull-request:
  block: asynchronously stop the VM on I/O errors
  vl: allow other threads to do qemu_system_vmstop_request
  sheepdog: fix NULL dereference in sd_create
  QemuOpts: check NULL opts in qemu_opt_get functions
  block: m25p80: Support read only bdrvs.
  block: m25p80: sync_page(): Deindent function body.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/mcayland/qemu-sparc' into staging

* remotes/mcayland/qemu-sparc:
  apb: Fix out-of-bounds array write access

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/mcayland/qemu-openbios' into staging

* remotes/mcayland/qemu-openbios:
  Update OpenBIOS images

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

console: move chardev declarations to sysemu/char.h

move generic chardev APIs to sysemu/char.h, to make them available to
callers which can not depend on the whole of ui/console.h.
This fixes a build error on systems without pixman-devel:

./configure --disable-tools --disable-docs --target-list=arm-linux-user
...
pixman            none
...
make
...
In file included from
/data/home/nchip/linaro/qemu/include/ui/console.h:4:0,
                 from /data/home/nchip/linaro/qemu/stubs/vc-init.c:2:
/data/home/nchip/linaro/qemu/include/ui/qemu-pixman.h:14:20: fatal
error: pixman.h: No such file or directory
 #include <pixman.h>
                    ^
compilation terminated.

Reported-by: Riku Voipio <riku.voipio@iki.fi>
Tested-by: Riku Voipio <riku.voipio@iki.fi>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Message-id: 1403508500-32691-1-git-send-email-mst@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

block: asynchronously stop the VM on I/O errors

With virtio-blk dataplane, I/O errors might occur while QEMU is
not in the main I/O thread.  However, it's invalid to call vm_stop
when we're neither in a VCPU thread nor in the main I/O thread,
even if we were to take the iothread mutex around it.

To avoid this problem, we can raise a request to the main I/O thread,
similar to what QEMU does when vm_stop is called from a CPU thread.
We know that bdrv_error_action is called from an AIO callback, and
the moment at which the callback will fire is not well-defined; it
depends on the moment at which the disk or OS finishes the operation,
which can happen at any time.  Note that QEMU is certainly not in a CPU
thread and we do not need to call cpu_stop_current() like vm_stop() does.

However, we need to ensure that any action taken by management will
result in correct detection of the error _and_ a running VM.  In particular:

- the event must be raised after the iostatus has been set, so that
"info block" will return an iostatus that matches the event.

- the VM must be stopped after the iostatus has been set, so that
"info block" will return an iostatus that matches the runstate.

The ordering between the STOP and BLOCK_IO_ERROR events is preserved;
BLOCK_IO_ERROR is documented to come first.

This makes bdrv_error_action() thread safe (assuming QMP events are,
which is attacked by a separate series).

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

vl: allow other threads to do qemu_system_vmstop_request

There patch protects vmstop_requested with a lock and introduces
qemu_system_vmstop_request_prepare.

Together with the new call to qemu_vmstop_requested in vm_start,
qemu_system_vmstop_request_prepare avoids a race where the VM could remain
stopped even though the iostatus of a block device has already been set
(for example).

qemu_system_vmstop_request_prepare however also lets the caller thread
delay observation of the state change until it has itself communicated
that change to the user.  This delay avoids any possibility of a wrong
reordering of the BLOCK_IO_ERROR event and the subsequent STOP event.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

sheepdog: fix NULL dereference in sd_create

Following command

qemu-img create -f qcow2 sheepdog:test 20g

will cause core dump because aio_context is NULL in sd_create. We should
initialize it by qemu_get_aio_context() to avoid NULL dereference.

Cc: qemu-devel@nongnu.org
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Liu Yuan <namei.unix@gmail.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

QemuOpts: check NULL opts in qemu_opt_get functions

Some places will call bdrv_create_file(filename, NULL, &local_err), where
opts is NULL. Check NULL in qemu_opt_get and qemu_opt_get_*_del functions,
to avoid extra effort of checking opts before calling them every time.

Signed-off-by: Chunyan Liu <cyliu@suse.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

block: m25p80: Support read only bdrvs.

By just never doing write-backs. This is completely invisible to the
guest, as the entire storage area is implemented as device state (at
realize time the entire drive is read in).

Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

block: m25p80: sync_page(): Deindent function body.

sync_page() was conditionalizing it's whole fn body on the bdrv being
non-null. Just return for the function immediately on NULL brdv and
get rid of the big if.

Makes implementation consistent with flash_zynq_area().

Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Update OpenBIOS images

Update OpenBIOS images to SVN r1306 built from submodule.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>

apb: Fix out-of-bounds array write access

The array regs is declared with IOMMU_NREGS (3) elements and accessed
using IOMMU_CTRL (0) and IOMMU_BASE (8). In most cases, those values
are right shifted before being used as an index which results in indices
0 and 1. In one case, this right shift was missing for IOMMU_BASE which
results in an out-of-bounds write access with index 8.

The patch adds the missing shift operation also for IOMMU_CTRL where
it is needed only for cosmetic reasons.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>

gt64xxx_pci: Add VMStateDescription

Add VMStateDescription for GT64120 PCI emulation used by the Malta
platform, to allow it to work with savevm/loadvm and live migration.

The entire register array is saved/restored using VMSTATE_UINT32_ARRAY
(fixed length GT_REGS = 1024).

Signed-off-by: Sanjay Lal <sanjayl@kymasys.com>
[james.hogan@imgtec.com: Convert to VMState]
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Andreas Färber <afaerber@suse.de>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>

target-mips: copy CP0_Config1 into DisasContext

In order to avoid access to the CPUMIPSState structure in the
translator, keep a copy of CP0_Config1 into DisasContext. The whole
register is read-only so it can be copied as a single value.

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>

Merge remote-tracking branch 'remotes/kvm/uq/master' into staging

* remotes/kvm/uq/master:
  hw/mips: malta: Don't boot from flash with KVM T&E
  MAINTAINERS: Add entry for MIPS KVM
  target-mips: Enable KVM support in build system
  hw/mips: malta: Add KVM support
  hw/mips: In KVM mode, inject IRQ2 (I/O) interrupts via ioctls
  target-mips: Call kvm_mips_reset_vcpu() from mips_cpu_reset()
  target-mips: kvm: Add main KVM support for MIPS
  kvm: Allow arch to set sigmask length
  target-mips: get_physical_address: Add KVM awareness
  target-mips: get_physical_address: Add defines for segment bases
  hw/mips: Add API to convert KVM guest KSEG0 <-> GPA
  hw/mips/cputimer: Don't start periodic timer in KVM mode
  target-mips: Reset CPU timer consistently
  KVM: Fix GSI number space limit

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

pc,pci,virtio,hotplug fixes, enhancements

numa work by Hu Tao and others
memory hotplug by Igor
vhost-user by Nikolay, Antonios and others
guest virtio announcements by Jason
qtest fixes by Sergey
qdev hotplug fixes by Paolo
misc other fixes mostly by myself

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
* remotes/mst/tags/for_upstream: (109 commits)
  numa: use RAM_ADDR_FMT with ram_addr_t
  qapi/string-output-visitor: fix bugs
  tests: simplify code
  qapi: fix input visitor bugs
  acpi: rephrase comment
  qmp: add ACPI_DEVICE_OST event handling
  qmp: add query-acpi-ospm-status command
  acpi: implement ospm_status() method for PIIX4/ICH9_LPC devices
  acpi: introduce TYPE_ACPI_DEVICE_IF interface
  qmp: add query-memory-devices command
  numa: handle mmaped memory allocation failure correctly
  pc: acpi: do not hardcode preprocessor
  qmp: clean out whitespace
  qdev: recursively unrealize devices when unrealizing bus
  qdev: reorganize error reporting in bus_set_realized
  qapi: fix build on glib < 2.28
  qapi: make string output visitor parse int list
  qapi: make string input visitor parse int list
  tests: fix memory leak in test of string input visitor
  hmp: add info memdev
  ...

Conflicts:
include/hw/i386/pc.h
[PMM: fixed minor conflict in pc.h]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20140619' into staging

target-arm:
 * Support PSCI 0.2 when using KVM
 * fix AIRCR reset value for v7M CPUs
 * report correct size information for pflash_cfi01
 * minor coverity fixes
 * avoid warnings on Windows builds due to #define clash
 * implement TTBCR PD0/PD1 bits

# gpg: Signature made Thu 19 Jun 2014 18:35:06 BST using RSA key ID 14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"

* remotes/pmaydell/tags/pull-target-arm-20140619:
  armv7m_nvic: fix AIRCR implementation
  Use PSCI v0.2 compatible string when KVM or TCG provides it
  target-arm: Introduce per-CPU field for PSCI version
  target-arm: Implement kvm_arch_reset_vcpu() for KVM ARM64
  target-arm: Enable KVM_ARM_VCPU_PSCI_0_2 feature when possible
  target-arm: Common kvm_arm_vcpu_init() for KVM ARM and KVM ARM64
  kvm: Handle exit reason KVM_EXIT_SYSTEM_EVENT
  hw/block/pflash_cfi01: Report correct size info for parallel configs
  hw/arm/vexpress: Forbid specifying flash contents in two ways at once
  target-arm/translate-a64.c: Fix dead ?: in handle_simd_shift_fpint_conv()
  target-arm/translate-a64.c: Remove dead ?: in disas_simd_3same_int()
  target-arm: Add ULL suffix to calculation of page size
  hw/arm/spitz: Avoid clash with Windows header symbol MOD_SHIFT
  target-arm: implement PD0/PD1 bits for TTBCR

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/kraxel/tags/pull-vnc-20140619-1' into staging

vnc: cleanups and fixes

# gpg: Signature made Thu 19 Jun 2014 12:02:09 BST using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"

* remotes/kraxel/tags/pull-vnc-20140619-1:
  vnc: fix screen updates
  vnc: Drop superfluous conditionals around g_strdup()
  vnc: Drop superfluous conditionals around g_free()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

spice: fix 32bit build

Tested-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1403244764-8622-1-git-send-email-kraxel@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/rth/tcg-next' into staging

* remotes/rth/tcg-next:
  tcg/optimize: Don't special case TCG_OPF_CALL_CLOBBER

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/mips: malta: Don't boot from flash with KVM T&E

In KVM trap & emulate (T&E) mode the flash reset region at 0xbfc00000
isn't executable, which is why the minimal kernel bootloader is loaded
and executed from the last 1MB of DRAM instead.

Therefore if no kernel is provided on the command line and KVM is
enabled, exit with an error since booting from flash will fail.

Reported-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

armv7m_nvic: fix AIRCR implementation

The returned reset value was wrong (off by one zero nibble), and
qemu didn't log unimplemented writes to the PRIGROUP field.

Signed-off-by: Oran Avraham <oranav@gmail.com>
Message-id: 1403010447-4627-1-git-send-email-oranav@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Use PSCI v0.2 compatible string when KVM or TCG provides it

If we have PSCI v0.2 emulation available for KVM ARM/ARM64 or TCG then
we need to provide PSCI v0.2 compatible string via generated DTB.

Signed-off-by: Pranavkumar Sawargaonkar <pranavkumar@linaro.org>
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Reviewed-by: Rob Herring <rob.herring@linaro.org>
Message-id: 1402901605-24551-9-git-send-email-pranavkumar@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target-arm: Introduce per-CPU field for PSCI version

We require to know the PSCI version available to given CPU at
potentially many places. Currently, we need to know PSCI version
when generating DTB for virt machine.

This patch introduce per-CPU 32bit field representing the PSCI
version available to the CPU. The encoding of this 32bit field
is same as described in PSCI v0.2 spec.

Signed-off-by: Pranavkumar Sawargaonkar <pranavkumar@linaro.org>
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1402901605-24551-8-git-send-email-pranavkumar@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target-arm: Implement kvm_arch_reset_vcpu() for KVM ARM64

To implement kvm_arch_reset_vcpu(), we simply re-init the VCPU
using kvm_arm_vcpu_init() so that all registers of VCPU are set
to their reset values by in-kernel KVM code.

Signed-off-by: Pranavkumar Sawargaonkar <pranavkumar@linaro.org>
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1402901605-24551-7-git-send-email-pranavkumar@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target-arm: Enable KVM_ARM_VCPU_PSCI_0_2 feature when possible

Latest linux kernel supports in-kernel emulation of PSCI v0.2 but
to enable it we need to select KVM_ARM_VCPU_PSCI_0_2 feature using
KVM_ARM_VCPU_INIT ioctl.

Also, we can use KVM_ARM_VCPU_PSCI_0_2 feature for VCPU only when
linux kernel has KVM_CAP_ARM_PSCI_0_2 capability.

This patch updates kvm_arch_init_vcpu() to enable KVM_ARM_VCPU_PSCI_0_2
feature for VCPU when KVM ARM/ARM64 has KVM_CAP_ARM_PSCI_0_2 capability.

Signed-off-by: Pranavkumar Sawargaonkar <pranavkumar@linaro.org>
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1402901605-24551-6-git-send-email-pranavkumar@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target-arm: Common kvm_arm_vcpu_init() for KVM ARM and KVM ARM64

Introduce a common kvm_arm_vcpu_init() for doing KVM_ARM_VCPU_INIT
ioctl in KVM ARM and KVM ARM64. This also helps us factor-out few
common code lines from kvm_arch_init_vcpu() for KVM ARM/ARM64.

Signed-off-by: Pranavkumar Sawargaonkar <pranavkumar@linaro.org>
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1402901605-24551-5-git-send-email-pranavkumar@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

kvm: Handle exit reason KVM_EXIT_SYSTEM_EVENT

In-kernel PSCI v0.2 emulation of KVM ARM/ARM64 forwards SYSTEM_OFF
and SYSTEM_RESET function calls to QEMU using KVM_EXIT_SYSTEM_EVENT
exit reason.

This patch updates kvm_cpu_exec() to handle KVM_SYSTEM_EVENT_SHUTDOWN
and KVM_SYSTEM_EVENT_RESET system-level events from QEMU-side.

Signed-off-by: Pranavkumar Sawargaonkar <pranavkumar@linaro.org>
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1402901605-24551-4-git-send-email-pranavkumar@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/block/pflash_cfi01: Report correct size info for parallel configs

If the flash device is configured with a device-width which is
not equal to the bank-width, indicating that it is actually several
narrow flash devices in parallel, the CFI table should report the
number of blocks and the size of a single device, not of the whole
combined setup. This stops Linux from complaining:
"NOR chip too large to fit in mapping. Attempting to cope..."

As usual, we retain the old broken but backwards compatible behaviour
when the device-width is not specified.

Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1402409025-25694-1-git-send-email-peter.maydell@linaro.org

hw/arm/vexpress: Forbid specifying flash contents in two ways at once

Detect attempts by the user to specify the contents of the first flash
device via both -bios and -drive if=pflash... simultaneously and
print a helpful error message.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1402419834-25982-1-git-send-email-peter.maydell@linaro.org

target-arm/translate-a64.c: Fix dead ?: in handle_simd_shift_fpint_conv()

In handle_simd_shift_fpint_conv(), the combination of is_double == true,
is_scalar == false and is_q == false is an unallocated encoding; the
'both parts false' case of the nested ?: expression for calculating
maxpass is therefore unreachable and can be removed.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: 1402171881-14343-4-git-send-email-peter.maydell@linaro.org

target-arm/translate-a64.c: Remove dead ?: in disas_simd_3same_int()

In disas_simd_3same_int(), none of the instructions permit is_q
to be false with size == 3 (this would be a vector operation with
a one-element vector, and the instruction set encodes those as
scalar operations). Replace the always-true ?: check with an
assert.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: 1402171881-14343-3-git-send-email-peter.maydell@linaro.org

target-arm: Add ULL suffix to calculation of page size

The maximum block size for AArch64 address translation is 2GB. This means
that we need a ULL suffix on our shift to avoid shifting into the sign
bit of a signed 32 bit integer.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: 1402171881-14343-2-git-send-email-peter.maydell@linaro.org

hw/arm/spitz: Avoid clash with Windows header symbol MOD_SHIFT

The Windows headers provided by MinGW define MOD_SHIFT. Avoid
it by using SPITZ_MOD_* for our constants here.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target-arm: implement PD0/PD1 bits for TTBCR

Corrected handling of writes to TTBCR for ARMv8 (previously UNK/SBZP
bits are not RES0) and ARMv7 (new bits PD0/PD1 for CPUs with Security
Extensions).

Bits PD0/PD1 are now respected in get_phys_addr_v6/v5() and
get_level1_table_address.

Signed-off-by: Fabian Aggeler <aggelerf@ethz.ch>
Message-id: 1402409556-18574-1-git-send-email-aggelerf@ethz.ch
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

numa: use RAM_ADDR_FMT with ram_addr_t

commit 4407ab055be995e64633322a78e64dfa376dc534
    vl.c: extend -m option to support options for memory hotplug
prints ram_addr_t with u64 format, this is wrong for
some systems, in particular w32.

print ram_addr_t with RAM_ADDR_FMT to fix build on w32.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

qapi/string-output-visitor: fix bugs

in human mode, we are creating the string:

16-31 (16-31)

instead of

16-17 (10-1f)

because we forgot to pass 'true' as the human parameter on one of the
two calls to format_string.
Also, this is a worsening of quality; previously we would produce

16 (0x10)

to make it obvious which number was hex.
Fix these issues.

Reported-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>

tests: simplify code

Use error_abort instead of open-coded assert.
Cleaner and shorter.

Reported-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>

qapi: fix input visitor bugs

Remove dead code.  Reset errno to 0 before each strtoull call, as the
man page requires.

Reported-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>

acpi: rephrase comment

"only upto" is not proper English.
Say "up to" and drop "only".

Reported-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

qmp: add ACPI_DEVICE_OST event handling

emits event when ACPI OSPM evaluates _OST method
of ACPI device.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

qmp: add query-acpi-ospm-status command

... to get ACPI OSPM status reported by ACPI devices
via _OST method.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

acpi: implement ospm_status() method for PIIX4/ICH9_LPC devices

... using TYPE_ACPI_DEVICE_IF interface.
Which provides status reporting of ACPI declared memory devices

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

acpi: introduce TYPE_ACPI_DEVICE_IF interface

... it will be used to abstract generic ACPI bits from
device that implements ACPI interface.

ACPIOSTInfo type is used for passing-through raw _OST
event/status codes reported by guest OS to a management
layer. It lets management tools interpret values
as specified by ACPI spec if it is interested in it.

QEMU doesn't encode these values as enum, since it
doesn't need to handle them and it allows interface
to scale well without any changes in QEMU while guest
OS and management evolves in time.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

qmp: add query-memory-devices command

... allowing to get state of present memory devices.
Currently implemented only for PCDIMMDevice.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

numa: handle mmaped memory allocation failure correctly

when memory_region_init_ram_from_file() fails
memory_region_size() will still return size that was
provided at region init time.
Instead use errp to properly detect error condition.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pc: acpi: do not hardcode preprocessor

but use one provided by environment, in addition
force C style preprocessing so that 'gcc -E' or
"clang -E" wouldn't ignore .dsl files.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

qmp: clean out whitespace

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

qdev: recursively unrealize devices when unrealizing bus

When the patch was posted that became 5c21ce7 (qdev: Realize buses
on device realization, 2014-03-12), it included recursive realization
and unrealization of devices when the bus's "realized" property
was toggled.

However, due to the same old worries about recursive realization
and prerequisites not being realized yet, those hunks were dropped when
committing the patch.  Unfortunately, this causes a use-after-free bug
(easily reproduced by a PCI hot-unplug action).

Before the patch, device_unparent behaved as follows:

   for each child bus
     unparent bus ----------------------------.
     | for each child device                  |
     |   unparent device ---------------.     |
     |   | unrealize device             |     |
     |   | call dc->unparent            |     |
     |   '-------------------------------     |
     '----------------------------------------'
   unrealize device

After the patch, it behaves as follows instead:

   unrealize device --------------------.
   | for each child bus                 |
   |   unrealize bus               (A)  |
   '------------------------------------'
   for each child bus
     unparent bus ----------------------.
     | for each child device            |
     |   unrealize device          (B)  |
     |   call dc->unparent              |
     '----------------------------------'

At the step marked (B) the device might use data from the bus that is
not available anymore due to step (A).

To fix this, we need to unrealize devices before step (A).  To sidestep
concerns about recursive realization, only do recursive unrealization
and leave the "value && !bus->realized" case as it is.

The resulting flow is:

   for each child bus
     unrealize bus ---------------------.
     | for each child device            |
     |   unrealize device          (B)  |
     | call bc->unrealize          (A)  |
     '----------------------------------'
   unrealize device
   for each child bus
     unparent bus ----------------------.
     | for each child device            |
     |   unparent device                |
     '----------------------------------'

where everything is "powered down" before it is unassembled.

Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Andreas Färber <afaerber@suse.de>

qdev: reorganize error reporting in bus_set_realized

No semantic change.

Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Andreas Färber <afaerber@suse.de>

qapi: fix build on glib < 2.28

The following commits:
    qapi: make string output visitor parse int list
    qapi: make string input visitor parse int list
break with glib < 2.28 since they use the
new g_list_free_full function.

Open-code that to fix build on old systems.

Cc: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

qapi: make string output visitor parse int list

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
MST: split up patch

qapi: make string input visitor parse int list

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
MST: split up patch

tests: fix memory leak in test of string input visitor

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hmp: add info memdev

This is the hmp counterpart of qmp query-memdev.

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
MST: fix build on 32 bit

qmp: add query-memdev

Add qmp command query-memdev to query for information
of memory devices

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hostmem: add properties for NUMA memory policy

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
[Raise errors on setting properties if !CONFIG_NUMA.  Add BUILD_BUG_ON
 checks. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hostmem: add property to map memory with MAP_SHARED

A new "share" property can be used with the "memory-file" backend to
map memory with MAP_SHARED instead of MAP_PRIVATE.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hostmem: allow preallocation of any memory region

And allow preallocation of file-based memory even without -mem-prealloc.
Some care is necessary because -mem-prealloc does not allow disabling
preallocation for hostmem-file.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hostmem: add merge and dump properties

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

osdep: add merge and dump flags

will be used by follow up patch

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hostmem: add file-based HostMemoryBackend

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
MST: comment tweak

hostmem: separate allocation from UserCreatable complete method

This allows the superclass to set various policies on the memory
region that the subclass creates. Drops hostmem-ram's complete method
accordingly.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

backend:hostmem: replace hostmemory with host_memory

..to keep names consistant.

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pc: pass MachineState to pc_memory_init

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Reviewed-By: Igor Mammedov <imammedo@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vl: redo -object parsing

Follow the lines of the HMP implementation, using OptsVisitor
to parse the options.  This gives access to OptsVisitor's
rich parsing of integer lists.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

memory: add error propagation to file-based RAM allocation

Right now, -mem-path will fall back to RAM-based allocation in some
cases.  This should never happen with "-object memory-file", prepare
the code by adding correct error propagation.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
MST: drop \n at end of error messages

memory: move mem_path handling to memory_region_allocate_system_memory

Like the previous patch did in exec.c, split memory_region_init_ram and
memory_region_init_ram_from_file, and push mem_path one step further up.
Other RAM regions than system memory will now be backed by regular RAM.

Also, boards that do not use memory_region_allocate_system_memory will
not support -mem-path anymore.  This can be changed before the patches
are merged by migrating boards to use the function.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

numa: add -numa node,memdev= option

This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes.  For example:

 -object memory-ram,size=1024M,policy=bind,host-nodes=0,id=ram-node0 \
 -numa node,nodeid=0,cpus=0,memdev=ram-node0 \
 -object memory-ram,size=1024M,policy=interleave,host-nodes=1-3,id=ram-node1 \
 -numa node,nodeid=1,cpus=1,memdev=ram-node1

The option replaces "-numa node,mem=".

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution

qom: introduce object_property_get_enum and object_property_get_uint16List

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Introduce signed range.

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
MST: split up patch

configure: add Linux libnuma detection

Add detection of libnuma (mostly contained in the numactl package)
to the configure script. Can be enabled or disabled on the command
line, default is use if available.

Signed-off-by: Andre Przywara <andre.przywara@amd.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

memory: move RAM_PREALLOC_MASK to exec.c, rename

Prepare for adding more flags.  The "_MASK" suffix is unique, kill it.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

memory: move preallocation code out of exec.c

So that backends can use it.

Since we need the page size for efficiency, move code to compute it
out of translate-all.c and into util/oslib-win32.c.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

memory: reorganize file-based allocation

Split the internal interface in exec.c to a separate function, and
push the check on mem_path up to memory_region_init_ram.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

numa: introduce memory_region_allocate_system_memory

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
MST: resolve conflicts

qmp: improve error reporting for -object and object-add

Use QERR_INVALID_PARAMETER_VALUE for consistency.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

man: improve -numa doc

The -numa option documentation in qemu's manpage lacks the command-line
options and some information regarding how it relates to options -m and
-smp. This commit fills in the missing text.

Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

NUMA: expand MAX_NODES from 64 to 128

libnuma choosed 128 for MAX_NODES, so we follow libnuma here.

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

NUMA: convert -numa option to use OptsVisitor

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Tested-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

NUMA: Add numa_info structure to contain numa nodes info

Add the numa_info structure to contain the numa nodes memory,
VCPUs information and the future added numa nodes host memory
policies.

Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Andre Przywara <andre.przywara@amd.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
[Fix hw/ppc/spapr.c - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

NUMA: check if the total numa memory size is equal to ram_size

If the total number of the assigned numa nodes memory is not
equal to the assigned ram size, it will write the wrong data
to ACPI table, then the guest will ignore the wrong ACPI table
and recognize all memory to one node. It's buggy, we should
check it to ensure that we write the right data to ACPI table.

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: error message reworded

NUMA: move numa related code to new file numa.c

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Andre Przywara <andre.przywara@amd.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: comment tweaks

tests: disable vhost test temporarily

This test needs a bit more work: issues have been
found on legacy systems, disable it for now to
avoid false positives for people.
Will re-enable after issues are addressed.

Reported-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Add qtest for vhost-user

This test creates a 'server' chardev to listen for vhost-user messages.
Once VHOST_USER_SET_MEM_TABLE is received it mmaps each received region,
and read 1k bytes from it. The read data is compared to data from readl.

The test requires hugetlbfs to be already mounted and writable. The mount
point defaults to '/hugetlbfs' and can be specified via the environment
variable QTEST_HUGETLBFS_PATH.

The rom pc-bios/pxe-virtio.rom is used to instantiate a virtio pcicontroller.

Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
MST: fix up coding style
MST: disable vhost test temporarily

This test needs a bit more work: issues have been
found on legacy systems, disable it for now to
avoid false positives for people.
Will re-enable after issues are addressed.

Reported-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

libqemustub: add stubs to be able to use qemu-char.c

chardev depends on lots of external symbols that are not necessarily
needed to be able to use, for example, 'socket chardev'. So add stubs
for these functions:

 - bdrv_commit_all
 - qemu_chr_open_msmouse
 - is_daemonized
 - qemu_add_machine_init_done_notifier
 - monitor_init
 - qemu_notify_event
 - vc_init

and this array:

 - serial_hds

Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Add vhost-user protocol documentation

This document describes the basic message format used by vhost-user
for communication over a unix domain socket. The protocol is based
on the existing ioctl interface used for the kernel version of vhost.

Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Add the vhost-user netdev backend to the command line

The supplied chardev id will be inspected for supported options. Only
a socket backend, with a set path (i.e. a Unix socket) and optionally
the server parameter set, will be allowed. Other options (nowait, telnet)
will make the chardev unusable and the netdev will not be initialised.

Additional checks for validity:
  - requires `-numa node,memdev=..`
  - requires `-device virtio-net-*`

The `vhostforce` option is used to force vhost-net when we deal with
non-MSIX guests.

Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>

Merge remote-tracking branch 'remotes/bonzini/scsi-next' into staging

* remotes/bonzini/scsi-next:
  virtio-scsi: define dummy handle_output for vhost-scsi vqs
  block/iscsi: drop obsolete pointers from iscsi_co_writev
  block/iscsi: fix init value for iTask->retries
  block/iscsi: bump libiscsi requirement to 1.9.0
  virtio-scsi: add support for the any_layout feature
  virtio-scsi: introduce virtio_scsi_complete_cmd_req
  virtio-scsi: prepare sense data handling for any_layout
  virtio-scsi: add extra argument and return type to qemu_sgl_concat
  virtio-scsi: add target swap for VirtIOSCSICtrlTMFReq fields
  virtio-scsi: start preparing for any_layout
  util: add return value to qemu_iovec_concat_iov
  megasas: use PCI DMA API
  scsi: Print command name in debug
  scsi-disk: fix bug in scsi_block_new_request() introduced by commit 137745c
  scsi-disk.c: Fix compilation with -DDEBUG_SCSI
  block/iscsi: use 16 byte CDBs only when necessary
  block/iscsi: fix potential segfault on early callback
  block/iscsi: handle BUSY condition

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

include/qemu/aes.h: Avoid conflicts with FreeBSD AES functions

FreeBSD's libcrypto provides functions with the same names as us;
use #define to rename our versions to avoid conflicts at link time.

Signed-off-by: Sean Bruno <sbruno@freebsd.org>
Signed-off-by: Ed Maste <emaste@freebsd.org>
Message-id: 1402930927-41125-1-git-send-email-sbruno@freebsd.org
[PMM: improved commit message, fixed comment typo]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

w32: Fix regression caused by new g_poll implementation

Commit 5a007547df76446ab891df93ebc55749716609bf tried to fix a
performance degradation caused by bad handling of small timeouts
in the original implementation of g_poll.

Since that commit, hard disk I/O no longer works.

Instead of rewriting the g_poll implementation, this patch simply copies
the original code (released under LGPL) from latest glib and only modifies
it where needed (see comments in the code). URL of the original code:
https://git.gnome.org/browse/glib/tree/glib/gpoll.c

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Message-id: 1401291744-14314-1-git-send-email-sw@weilnetz.de
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Add new vhost-user netdev backend

Add a new QEMU netdev backend that is intended to invoke vhost_net with the
vhost-user backend. It uses an Unix socket chardev to establish a
communication with the 'slave' (client and server mode supported).

At runtime the netdev will handle OPEN/CLOSE events from the chardev. Upon
disconnection it will set link_down accordingly and notify virtio-net; the
virtio-net interface will go down.

Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-net: vhost-user feature bits support

Handle the feature bits negotiation when using vhost-user. Allow
the underlying implementation to have a finer control over all the
bits except the VIRTIO_NET_F_MAC.

Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Add vhost-user as a vhost backend.

The initialization takes a chardev backed by a unix domain socket.
It should implement qemu_fe_set_msgfds in order to be able to pass
file descriptors to the remote process.

Each ioctl request of vhost-kernel has a vhost-user message equivalent,
which is sent over the control socket.

The general approach is to copy the data from the supplied argument
pointer to a designated field in the message. If a file descriptor is
to be passed it will be placed in the fds array for inclusion in
the sendmsg control header.

VHOST_SET_MEM_TABLE ignores the supplied vhost_memory structure and scans
the global ram_list for ram blocks with a valid fd field set. This would
be set when the '-object memory-file' option with share=on property is used.

Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Add vhost-backend and VhostBackendType

Use vhost_set_backend_type to initialise a proper vhost_ops structure.
In vhost_net_init and vhost_net_start_one call conditionally TAP related
initialisation depending on the vhost backend type.

Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>